Part Five — Document Projects Chapter Twenty-one

Designing a Long-Form Report

A report is not a book that failed to become a novel. It has its own conventions, its own hierarchy of readers, and its own obligation to move between argument and evidence in ways that literary prose does not.

A long-form report is read differently from a book. Readers of a book begin at page one and proceed forward. Readers of a report arrive at different entry points — the executive summary, the section most relevant to their brief, a specific figure cited in a meeting — and navigate laterally across sections rather than linearly through them. The design of a report must accommodate both kinds of reading: the executive who reads only the first six pages, the analyst who reads every table, the policymaker who searches for a specific recommendation and needs to find it quickly. A report that is designed only for sequential reading will fail its busiest readers. A report designed only for navigation will feel thin and graceless to anyone who reads it properly. The design problem is to serve both.

This chapter builds a long-form report from the ground up, introducing each major element — cover, executive summary, table of contents, body sections, data tables, callout statistics, sidebars — with both its design rationale and its CSS implementation. The report we build is a generic but realistic 30-page policy document: the kind of thing produced by research institutes, government agencies, and professional services organisations for audiences that are expert in the subject but time-constrained in their reading.

· · ·

How a report differs from a book Structure, hierarchy, and the impatient reader

The most important structural difference between a report and a book is the hierarchy of readers. A book has, in effect, one reader: someone who will read from beginning to end. A report has several simultaneous readers at different depths of engagement, and the design must serve all of them. The executive who has twelve minutes reads the cover, the executive summary, and the key findings. The section lead reads the introduction, the relevant body sections, and the conclusion. The researcher reads everything, including the appendices, notes, and methodology. The design must make each of these reading paths clear, efficient, and self-contained without making any of them feel like a degraded version of the full experience.

This creates a hierarchy of typographic emphasis quite different from a book's. Where a book's hierarchy is primarily sequential — one chapter leading to the next, subheadings marking progress through an argument — a report's hierarchy is primarily navigational. The section numbers, the contents list, the running headers, the pull-out statistics, the clearly labelled figures: all of these exist to help readers locate, skip, and return to material according to their specific needs. A report with no section numbers is a report that makes navigation harder than it needs to be. A report with no pull-out statistics makes the executive summary reader do more work than necessary.

The other major structural difference is the relationship between prose and non-prose content. In a book, prose is primary and tables, figures, and illustrations are supplementary. In a report, this relationship is often reversed: the tables and figures may be the primary evidence, with the prose serving as interpretation and context. This means the design must give visual prominence to data — not by making it decorative, but by giving it the space, the clarity, and the typographic distinction it needs to be read accurately and quickly.

Report anatomy — the six zones of a long-form report
Cover EXECUTIVE SUMMARY Contents BODY SECTIONS 47% Appendices · Bibliography brand / identity executive reads this navigation analyst reads this — prose + data + charts researcher reads this

Figure 21.1 — The six zones of a long-form report. Cover (brand identity, sets register); executive summary (key findings in condensed form, the only section many decision-makers read); table of contents (navigational infrastructure); body sections (the full argument — prose, data tables, charts, sidebars); appendices and bibliography (supporting material for researchers and verifiers). The design must serve each zone's different reading context while maintaining visual coherence across all of them.

· · ·

The executive summary Six pages that carry the whole argument

The executive summary is the most read section of any report, and frequently the only section read by the people with decision-making authority. It must be complete enough to be self-contained, concise enough to be read in fifteen minutes, and clear enough to be understood without reference to the full document. These requirements determine both its content — a tight synthesis of findings, conclusions, and recommendations — and its design.

The executive summary benefits from a distinct visual treatment that signals its special status within the document. The most effective approach is a dark or tinted background for the entire section — a visual container that makes the executive summary immediately identifiable when the document is opened at any page, and creates a clear boundary between the navigational front matter and the body that follows. When using a dark background, the text must be in a light color, and all typographic elements — headings, body text, key findings — must be re-specified against the dark ground rather than relying on the default light-background values.

Within the executive summary, key findings are often displayed as pull-out statistics or numbered conclusions. These give the busy reader visual anchors — the places they can look first and still extract the essential quantitative results. The design of these statistics matters: a very large number set in display type communicates scale and confidence; the same number in body-weight text communicates information without emphasis. Choose the treatment that matches the claim.

Executive summary — live example dark background + stat callouts
Executive Summary

Urban tree cover in mid-sized European cities declined by an average of 6.3 percentage points between 2005 and 2023, with losses concentrated in rapidly developing peripheral districts. The decline correlates strongly with increased surface temperatures, reduced biodiversity indices, and measurably higher summer energy consumption in affected neighbourhoods.

This report examines the causes of decline, the policy instruments that have proved effective in arresting it, and the projected costs and benefits of three intervention scenarios at the municipal level. The analysis draws on remote-sensing data for 47 cities, supplemented by case studies in Oslo, Valencia, and Poznań.

6.3% Average decline in urban tree cover, 2005–2023 Across 47 mid-sized European cities
2.1°C Additional summer heat in areas with ≥10% canopy loss Compared to matched districts retaining cover
€340M Estimated 10-year cost of restoration across study cities Scenario B — moderate intervention

Figure 21.2 — Executive summary with key statistics. A dark-background summary panel establishes the register and contains the condensed argument. Three callout statistics beneath it — each with a large display number, a descriptive label, and a context note — give the executive reader three quantitative anchors without requiring them to read the full table. The statistics use tabular lining numerals for vertical alignment of the digits.

/* ── Executive summary — dark panel ─────────────── */
.exec-summary {
  background:      var(--ink);
  color:           rgba(248, 245, 240, 0.88);
  padding:         1.75rem 2rem;
  page:            exec-summary;
  break-before:    page;
  break-inside:    avoid;
  /* Ensure background prints                     */
  print-color-adjust: exact;
  -webkit-print-color-adjust: exact;
}

/* ── Key statistics row ────────────────────────── */
.stat-row {
  display:         grid;
  grid-template-columns: repeat(3, 1fr);
  gap:             1px;
  break-inside:    avoid;
}

.stat-number {
  font-family:     var(--font-body);
  font-size:       var(--text-2xl);
  line-height:     1;
  font-feature-settings: "lnum" 1, "tnum" 1;
}
· · ·

Data tables Presenting numbers with precision and clarity

Tables in a report are not decoration. They are primary evidence, and they must be designed to be read — accurately, quickly, and without ambiguity. The principles of good table design are few and well established: aligned columns, consistent units, clear headers, minimal visual noise, and a clear hierarchy between data cells and summary or total rows.

The first typographic decision for a data table is the numeral style. Body text uses old-style proportional numerals for integration with prose — but data tables require lining tabular numerals, which are all the same height and all the same width. Same height so the numbers do not produce an irregular visual rhythm across the table; same width so numbers in the same column align vertically, allowing the eye to compare magnitudes at a glance. Set font-variant-numeric: lining-nums tabular-nums on the table element, or equivalent font-feature-settings as shown in Chapter 10.

Alignment is the second decision. Text columns are left-aligned; number columns are right-aligned. This is not an aesthetic preference — it is a functional requirement. Right-aligned numbers place all units digits in the same horizontal position, making it immediately apparent which number is larger. A column of right-aligned numbers can be compared top to bottom with minimal eye movement. A column of left-aligned numbers cannot, because the units digits may be anywhere in the column width depending on the number of digits.

Zebra striping — alternating row backgrounds — is optional but useful for wide tables where the eye might lose its row and track across to the wrong data. A subtle tint, just a few percent above the background white, is sufficient. Avoid strong alternating colors that compete with the data or create a distracting grid pattern.

Data table — urban tree cover by city font-variant-numeric: lining-nums tabular-nums
City Country Cover 2005 (%) Cover 2023 (%) Change (pp) Pop. density (km²)
Oslo Norway 38.4 41.2 +2.8 1,491
Valencia Spain 12.7 9.3 −3.4 5,312
Poznań Poland 27.1 22.8 −4.3 1,730
Ghent Belgium 18.9 16.2 −2.7 2,215
Tampere Finland 44.2 45.8 +1.6 687
Trieste Italy 15.3 11.1 −4.2 3,748
Study median (n = 47) 24.6 21.3 −3.3 2,184

Figure 21.3 — A data table with correct typographic treatment. Dark header row establishes column structure; alternating row tints reduce tracking errors across wide rows; numeric columns are right-aligned with tabular lining numerals; directional change values use colour to encode direction (green for gain, red for loss) without requiring the reader to parse sign characters; footer row carries the study median in a visually distinct treatment. The table is sized to fit the text column width without horizontal overflow.

· · ·

Sidebars and pull-out elements Serving the skimmer without shortchanging the reader

A sidebar is a block of text that runs parallel to the main text but is visually distinguished from it — by a border, a tint, a different typeface weight, or its position in a margin column. Sidebars serve readers who are skimming by providing self-contained summaries, definitions, or contextual information that does not interrupt the main argument. They serve readers who are reading deeply by providing supplementary material without forcing them to leave the text and find an endnote.

The design of a sidebar must establish its relationship to the main text clearly. Too visually similar, and it blurs into the prose. Too visually different, and it becomes a separate document that competes for attention. The middle ground — a light tint background or a left border rule, the same typeface but a slightly smaller size or a different weight, a label identifying it as supplementary — communicates supplementary without competing.

Pull quotes in a report serve a different function than in editorial design. Where a magazine pull quote is primarily a design device — a way to break up a text column visually — a report pull quote is a navigational device: it surfaces a key statement that the skimming reader should not miss. A report pull quote should be genuinely informative, not merely atmospheric. It should be a claim or conclusion that stands alone without context, not a sentence that only makes sense if the surrounding paragraphs have been read.

Sidebar alongside body text CSS Grid: text column + sidebar column
3.2 Policy interventions

Tree protection ordinances

Municipal ordinances that require permits for the removal of mature trees — typically defined as those with a trunk diameter exceeding 30cm at breast height — have proved the most cost-effective single intervention available to local authorities. The evidence from the Oslo and Tampere case studies suggests that permit requirements reduce net canopy loss by 60–70% relative to cities without such regulations, even in rapidly developing areas.

The critical design feature is the compensation requirement: where permit-holders are allowed to proceed with removal, they must replace the removed canopy area within three years, typically at a ratio of 3:1 — three replacement trees for every mature specimen removed. Cities that implement permits without compensation requirements see smaller reductions in net loss, suggesting that the financial burden of replacement is the mechanism, not the administrative friction of the permit process itself.

Key definition
A mature tree is defined for regulatory purposes as a specimen with a trunk diameter of 30cm or more at breast height (1.3m above ground). Smaller specimens are not covered by protection ordinances in most of the study cities, though some — notably Oslo — apply a size-independent approach based on species designation.
Case study
Oslo, 2011: Introduction of the 3:1 replacement ordinance reduced net annual canopy loss from 340 hectares to 92 hectares within four years. Compliance rate: 94% in years 1–4.

Figure 21.4 — Body text with adjacent sidebars. A two-column CSS Grid positions the main argument in the wider left column and supplementary sidebars in the narrower right column. The sidebars use the same accent-color left border and highlight tint as callout elements elsewhere in the document — visual consistency signals to the reader that these elements are always supplementary, never primary. The main text reads without interruption; the sidebar reader can engage or skip without losing their place.

· · ·

Section hierarchy and navigation Numbering, headings, and the table of contents

A long-form report requires a numbering system for its sections. This is not primarily a typographic requirement — it is a navigational and referential one. Section numbers allow readers to navigate by number rather than by title, allow the table of contents to map the document structure precisely, and allow external references to cite specific sections without ambiguity. A report that says "see the section on policy interventions" is less usable than one that says "see §3.2."

In HTML and CSS, section numbers can be generated automatically using CSS counters, which increment at each heading element and can be concatenated to produce hierarchical numbering (1.1, 1.2, 2.1, 2.2, and so on) without manual numbering in the HTML source. This is preferable to hardcoded numbers in the HTML, because content can be reordered without renumbering. Paged.js respects CSS counters in the same way as screen CSS:

/* ── Automatic section numbering ──────────────── */
body          { counter-reset: section; }
.report-section { counter-increment: section;
                   counter-reset: subsection; }
.report-subsection { counter-increment: subsection; }

/* Heading labels from counters                  */
.report-section h2::before {
  content: counter(section) ". ";
  color: var(--ink-faint);
  font-family: var(--font-ui);
  font-style: normal;
  font-size: var(--text-sm);
  letter-spacing: 0.04em;
}

.report-subsection h3::before {
  content: counter(section) "." counter(subsection) "  ";
  color: var(--ink-faint);
  font-family: var(--font-ui);
  font-style: normal;
}

/* Table of contents: use leader dots            */
.toc-entry {
  display: flex;
  justify-content: space-between;
  gap: 0.5rem;
  align-items: baseline;
}
.toc-leader {
  flex: 1;
  border-bottom: 1px dotted var(--rule);
  margin-bottom: 0.2rem;
  margin-left: 0.5rem;
  margin-right: 0.5rem;
}

The complete report stylesheet

The companion files for this chapter include a complete report stylesheet that implements all the elements described here: the dark executive summary panel, the stat callout row, the data table (with dark header, zebra striping, and footer row), the sidebar grid layout, and the automatic section-numbering counter system. It is built on the same custom property foundation as the book interior stylesheet from Chapter 20 and can be combined with or used independently of it.

The key difference from the book interior stylesheet is the grid structure: where the book uses a single text column with a narrow notes margin, the report uses a two-column grid in body sections — a wider left column for the main text and a narrower right column for sidebars, callout statistics, and figures. This layout can be activated or deactivated per section using a class on the section container, allowing sections that do not use sidebars to default to a single-column layout without structural changes to the HTML.

· · ·

A well-designed report respects all its readers at once: the executive who wants findings, the analyst who wants evidence, the researcher who wants methods. None of these readers is wrong. All of them have legitimate claims on the document's design. Serving them simultaneously — through clear hierarchy, honest data presentation, and consistent navigational structure — is what distinguishes a report that is used from one that is merely submitted.

Chapter 22 turns to a quite different challenge: the editorial layout. Where the report values clarity, efficiency, and hierarchy, the editorial layout values energy, visual contrast, and the interplay of type and image. The same underlying CSS system applies, but the design sensibility is entirely different.